1. Field of the Invention
The present invention relates to an image processing apparatus and image processing method, which form an output image based on an input image, and an image processing apparatus and program, which record an output image on a print recording medium such as a paper sheet or the like.
2. Description of the Related Art
In general, a digital copying machine scans a document image for respective pixels using an image sensor or the like, and A/D (analog-to-digital) converts an analog electrical signal obtained as the output of the image sensor. The digital copying machine applies various processes to the obtained digital signal, and supplies the processed signal to a recording device to record (to be also referred to as “print” hereinafter) an output image.
Especially, a digital color copying machine scans a document image printed or drawn on a paper sheet while color-separating it into R (red), G (green), and B (blue). Then, the digital color copying machine sets the obtained RGB digital data as document data, and decides the output amounts of color materials represented by C (cyan), M (magenta), and Y (yellow) and a monochrome material represented by K (black) based on the document data. After that, the digital color copying machine sets, e.g., the obtained CMYK digital data as output data, and outputs an output image to a print recording medium (to be referred to as a recording medium hereinafter) based on the output data.
When a document which includes both a photograph and text (for example, an article of a magazine or the like) is used as a document image, it is normally the case to form, in the document image, a photograph part mainly by a color image, and to form a text part mainly by a monochrome image. A case will be examined below wherein, for example, an output image is to be output onto a recording medium such as a paper sheet or the like upon scanning a monochrome document image to form an output image. In this case, as for a text part formed by monochrome image information in the document image, it is often desirable to output a corresponding part in the output image using only a monochrome material. The reasons for this will be described below.
First, the text part mainly includes monochrome images of characters. Therefore, upon recording these character monochrome images, only the monochrome material is used without mixing any color materials to obtain output images with sharp edges.
Second, since the output process is done using only the monochrome material for a monochrome image part, the data size to be handled can be reduced compared to that including the color materials. Therefore, computation processes can be simplified.
Third, assume that a printing device like an ink-jet printing device to be described later is used, which provides a mechanism that adopts, as an output system, a serial system that sequentially outputs a document image for respective lines, thereby performing an output process for respective lines. In this case, assume that a mechanism is adopted, which outputs a larger number of lines at once upon using only the monochrome material than a case including the color materials. Upon recording a document image on which a monochrome image part continues over a plurality of lines, a larger number of lines can be output per unit time compared to the remaining image part, thus speeding up recording.
However, upon scanning a document image including both a photograph and text to form an output image, a monochrome image part in the document image often requires the color materials. The reasons for this will be described below.
As the first reason, it is desirable to generate output data that does not require any color material upon outputting a blank part of a document when a document image of a monochrome image part is scanned using an image sensor or the like. However, the color of original medium itself such as a paper sheet or the like used as an original document (so-called a background color) is not white in practice. Therefore, output data that reflects the background color is generated, and as a result, even output data of the blank part of the document requires the color materials.
As the second reason, it is desirable to generate output data which uses only the monochrome material upon outputting parts corresponding to the black portions of a document when a monochrome document image is scanned using the image sensor or the like. However, the parts corresponding to the black portions of the document are not purely black and have a tint of some other color. For this reason, output data that reflects a color scanned by the image sensor is generated, and as a result, the output data of the part corresponding to black of the document requires the color materials.
As the third reason, the image sensor scans a part corresponding to a boundary between a blank part and black part of a document as an intermediate color between white and black depending on the quality of the document image and the scanning precision of the image sensor, when a document image of a monochrome image part is scanned using the image sensor or the like. Therefore, output data that reflects such intermediate color is generated, and as a result, the output data of the part corresponding to the boundary between the blank part and black part of the document requires the color materials.
In order to solve the problem posed due to the first reason, an under-color removal process can be applied, which modifies the scanned document image by converting pixels with high luminance values into white pixels, as described in, e.g., Japanese Patent No. 02605316.
In order to solve the problem posed due to the second reason, a black removal process can be applied, which modifies the scanned document image by converting pixels with low luminance values into black pixels in the same manner as the under-color removal process.
A combined effect of the under-color removal process and black removal process can be obtained by a tone correction process of an image, as disclosed in, e.g., Japanese Patent Laid-Open No. 05-091322.
Furthermore, in order to solve the problem posed due to the third reason, a modification process that removes an intermediate color has been proposed. As a conventional method of the modification process for removing an intermediate color, a technique that binarizes an image using an arbitrary threshold, and the like, has been proposed.
For example, as a prior art that combines these techniques and obtains a preferable binarization result by applying binarization after the tone correction, Japanese Patent Laid-Open No. 2001-045303 is available. As described in Japanese Patent Laid-Open No. 2001-045303, minimum and maximum densities of each channel are derived for density histograms of respective RGB components of a document image, and each density histogram distribution is expanded to have the derived minimum and maximum densities as values at the two ends of a data value range, thus implementing the binarization process.
However, upon applying the technique disclosed in Japanese Patent Laid-Open No. 2001-045303 to a color image, not only a black character area but also a color image area are binarized. Hence, even a document including both black characters and a color photograph is binarized to a monochrome image.
To take measures, to perform area segmentation described in the related arts (Japanese Patent Laid-Open No. 2000-316097, Japanese Patent Laid-Open No. 2001-223915, and the like) and to apply the binarization process to only a part judged as black characters can be thought of. However, since the binarization is turned on/off at a boundary between the part judged as black characters and the remaining part, image quality considerably changes upon switching areas, and high image quality cannot be obtained.
That is, the related arts can modify a monochrome image part to an image that does not require any color materials. A case will be examined below wherein a mechanism that adopts a serial recording system like an ink-jet printing device to be described later and outputs an image for respective lines by using a monochrome material alone upon recording an image that has undergone such modification. In this case, since the mechanism that can output a larger number of lines at once compared to a case including color materials is adopted, the recording speed of an image including black characters together can be increased. However, in such a case, image quality deteriorates considerably.
In order to obtain high image quality, a black character area is binarized to black and white, while image processes which continuously change according to feature amounts of image areas so as not to impair original tones are applied to a smooth tone area.